Federated learning device interworking with mobile core system and method thereof
The present invention related to a method for federated learning method interworking with a mobile core system, the method comprising: querying terminal information of each individual terminal among a plurality of terminals; querying network performance information; selecting participating terminals among the plurality of terminals on the basis of the terminal information and the network performance information; transmitting respective parameters to the participating terminals and requesting local learning; and integrating the parameters.
1. A federated learning method interworking with a mobile core system, the method comprising:
querying terminal information of each individual terminal among a plurality of terminals;
querying network performance information;
selecting participating terminals among the plurality of terminals on the basis of the terminal information and the network performance information;
transmitting respective parameters to the participating terminals and requesting local learning; and
integrating the parameters,
the method further comprising:
receiving, by a mobile core, a terminal information query request;
querying the terminal information of each individual terminal among the plurality of terminals;
querying the network performance information;
integrating the terminal information and the network performance information; and
transferring at least one of the integrated terminal information and the integrated network performance information to a parameter server,
wherein the integrating of the terminal information and the network performance information integrates the terminal information and the network performance information by using AISF (Artificial Intelligence support function).
2. The method of claim 1 , further comprising, after the selecting of the participating terminals:
requesting setting of respective routes between the parameter server and the participating terminals; and
requesting releasing of the respective routes between the parameter server and the participating terminals.
3. The method of claim 1 , further comprising, after the selecting of the participating terminals:
requesting respectively dedicated slices when conducting federated learning;
requesting respective slice connections between the parameter server and the participating terminals; and
requesting releasing of the slices and releasing of the slice connections between the parameter server and the participating terminals.
4. The method of claim 1 , wherein the querying of the terminal information of each individual terminal among the plurality of terminals queries at least one of performance, a device type, reliability, and resource information of each individual terminal.
5. The method of claim 1 , wherein the querying of the network performance information queries at least one of a network topology, the network performance information, access performance of each terminal, and connection stability information by network type.
6. The method of claim 1 , further comprising:
further selecting the participating terminals on the basis of at least one of the integrated terminal information and the integrated network performance information; and
transferring the list of the selected participating terminals to the parameter server.
7. The method of claim 1 , further comprising, after the selecting of the participating terminals:
setting at least one of respective routes and bandwidths between the parameter server and the participating terminals.
8. The method of claim 7 , wherein the setting of at least one of the respective routes and bandwidths between a parameter server and the participating terminals sets at least one of the respective routes and bandwidths between the parameter server and the participating terminals by using AISF (Artificial Intelligence support function).
9. A federated learning device interworking with a mobile core system, the device comprising:
a plurality of terminals;
a mobile core comprising terminal information and network performance information; and
a parameter server configured to query the terminal information of each individual terminal among the plurality of terminals, query the network performance information, select participating terminals among the plurality of terminals on the basis of the terminal information and the network performance information, transmit respective parameters to the participating terminals, request local learning, and integrate the parameters,
wherein the mobile core receives a terminal information query request, queries the terminal information of each individual terminal among the plurality of terminals, queries the network performance information, integrates the terminal information and the network performance information, and transfers at least one of the integrated terminal information and the integrated network performance information to the parameter server,
and wherein the mobile core sets at least one of the respective routes and bandwidths between the parameter server and the participating terminals by using AISF (Artificial Intelligence support function).
10. The device of claim 9 , wherein the parameter server requests setting of respective routes between the parameter server and the participating terminals, and requests releasing of the respective routes between the parameter server and the participating terminals.
11. The device of claim 9 , wherein when conducting federated learning, the parameter server requests respectively dedicated slices, requests respective slice connections between the parameter server and the participating terminals, and requests releasing of the slices and releasing of the slice connections between the parameter server and the participating terminals.
12. The device of claim 9 , wherein the parameter server queries at least one of performance, a device type, reliability, and resource information of each individual terminal.
13. The device of claim 9 , wherein the parameter server queries at least one of a network topology, the network performance information, access performance of each terminal, and connection stability information by network type.
14. The device of claim 9 , wherein the mobile core further selects the participating terminals on the basis of at least one of the integrated terminal information and the integrated network performance information, and transfers a list of the selected participating terminals to the parameter server.
15. The device of claim 9 , wherein the mobile core sets at least one of respective routes and bandwidths between the parameter server and the participating terminals.
16. The device of claim 15 , wherein the mobile core sets at least one of the respective routes and bandwidths between the parameter server and the participating terminals by using AISF (Artificial Intelligence support function).